Professional Documents
Culture Documents
Gis For Archaeology
Gis For Archaeology
July 2009
Table of Contents
What Is GIS?
13
19
25
29
33
39
What Is GIS?
Making decisions based on geography is basic to human thinking. Where shall we go, what will it be
like, and what shall we do when we get there are applied to the simple event of going to the store or
to the major event of launching a bathysphere into the ocean's depths. By understanding geography
and people's relationship to location, we can make informed decisions about the way we live on our
planet. A geographic information system (GIS) is a technological tool for comprehending geography
and making intelligent decisions.
GIS organizes geographic data so that a person reading a map can select data necessary for a
specific project or task. A thematic map has a table of contents that allows the reader to add layers
of information to a basemap of real-world locations. For example, a social analyst might use the
basemap of Eugene, Oregon, and select datasets from the U.S. Census Bureau to add data layers
to a map that shows residents' education levels, ages, and employment status. With an ability to
combine a variety of datasets in an infinite number of ways, GIS is a useful tool for nearly every field
of knowledge from archaeology to zoology.
A good GIS program is able to process geographic data from a variety of sources and integrate
it into a map project. Many countries have an abundance of geographic data for analysis, and
governments often make GIS datasets publicly available. Map file databases often come included
with GIS packages; others can be obtained from both commercial vendors and government
agencies. Some data is gathered in the field by global positioning units that attach a location
coordinate (latitude and longitude) to a feature such as a pump station.
GIS maps are interactive. On the computer screen, map users can scan a GIS map in any direction,
zoom in or out, and change the nature of the information contained in the map. They can choose
whether to see the roads, how many roads to see, and how roads should be depicted. Then
they can select what other items they wish to view alongside these roads such as storm drains,
gas lines, rare plants, or hospitals. Some GIS programs are designed to perform sophisticated
calculations for tracking storms or predicting erosion patterns. GIS applications can be embedded
into common activities such as verifying an address.
From routinely performing work-related tasks to scientifically exploring the complexities of our world,
GIS gives people the geographic advantage to become more productive, more aware, and more
responsive citizens of planet Earth.
WWW.ESRI.COM
Base GIS layer incorporating oil spill geographic response plan and watershed areas with tribal contact areas.
During an emergency response to such an oil spill, an effective response requires accurate
information and clear communications. To meet this criteria, the Washington State Department
GIS BEST PRACTICES
WWW.ESRI.COM
of Archaeology and Historic Preservation (DAHP) and its GIS consultant, GeoEngineers, Inc.,
have created a GIS application that provides the Oil Spill Incident Command Structure current
data on archaeological site locations during an oil spill. This application was developed for the
Washington State Department of Ecology Oil Spill Response Program and includes contact
information for notifying concerned tribal government cultural staff. This Oil Spill Tribal Contact
tool provides quick and easy access to current tribal contact names for geographic regions.
During a recent spill event, the ecology responders were able to quickly identify the tribes of
concern by using the application. By clicking on a map at the location of an oil spill, a list of
federally recognized tribes was generated on the fly through an associated database.
Dr. Allyson Brooks, State Historic Preservation officer and director of DAHP, states, "Our ability
to protect these unique and special places in Washington is directly related to the speed with
which we can get information to the tribes and other decision makers and first responders on
the scene. By providing current data in spatial and tabular format, GIS helps us accomplish that
mission."
DAHP uses a combination of ArcView and ArcInfo for digitizing and maintaining archaeology site
locations. As part of DAHP's GIS initiative, more than 19,000 archaeological sites have been
mapped in GIS and attributed with information regarding site type (e.g., burials, petroglyphs).
Each site is mapped within a geodatabase and linked to a scanned image of the original
documentation regarding the discovery of the site. This data is available to DAHP staff via
a customized ArcGIS Desktop interface. It is not subject to public disclosure; however, data
sharing is possible with governmental agencies through memorandums of understanding. In
this way, DAHP is able to share this information with the Washington Department of Ecology for
protection of these valuable resources during oil spill events.
With funding from the state's Coastal Protection Fund and the assistance of the Tribal Historic
Preservation Offices and cultural staff of the 29 tribes, DAHP created a series of specific GIS
layers for each tribal government reflecting its geographic area of interest, its reservation lands,
and watersheds for which each tribe's cultural department wanted notification when a spill
occurred.
JULY 2009
GIS IN ARCHAEOLOGY
In many parts of the state, multiple tribes require notification. Also along the Columbia and
Snake rivers, tribes now resident in Oregon and Idaho have an interest and require notification.
The GIS application has a database table identifying any archaeological site potentially at risk
and the specific cultural staff contact with phone number, an after-hours emergency number,
and details on any staff who have hazmat training. These screens were created for ease of
use during an emergency and allow for quick printing of contact sheets. Also, since names and
phone numbers frequently change, the table was created to allow for ease of updating.
(Reprinted from the Spring 2006 issue of ArcNews magazine)
WWW.ESRI.COM
A marvelous opportunity existed, it seemed, to wed history and technology, but the solution
had to be able to make maps, display photos, query databases, and allow for customization, all
in a Web environment. Since Boulder has been a longtime user of ArcGIS Desktop (ArcView,
ArcEditor, ArcInfo) and ArcSDE and knows the power of GIS to integrate information, ArcIMS
seemed to be the logical choice to bring this disparate data into a unique and informative Web
site.
The interactive map of Columbia Cemetery is an HTML viewer page created by ArcIMS. The user
can zoom in and out of the map and get information about who is buried in a particular cemetery lot.
Columbia Cemetery is a virtual "Who's Who" of early Bouldera historic, cultural, and
artistic resource containing the remains of many of the city's founders and pioneers. Initially
established in 1870 on 10 acres of cattle-grazed pastureland, the cemetery today has 6,500
JULY 2009
GIS IN ARCHAEOLOGY
burials and 3,000 headstones. Like many other Victorian era cemeteries, Columbia contains
grave markers of various types: monuments, crosses, obelisks, and tablets made of marble,
granite, sandstone, limestone, and wood. The tombstones not only mark the graves of early
pioneers who have helped make Boulder what it is today, but they are also narratives describing
Colorado's social and economic structure, its religious tenets, and ethnic composition. The
epitaphs, engravings, and decorations provide insight into earlier customs, religious beliefs,
folklore, art, and medicine. Homemade Depression-era "folk markers" are juxtaposed with
ornate and towering granite monuments belonging to bank presidents. Marble lambs and doves
mark the graves of children felled by scarlet fever, diphtheria, and tuberculosis, and graves
adorned with flowers, stuffed animals, and coins poignantly indicate recent visits to century-old
burials.
WWW.ESRI.COM
Columbia Cemetery is owned by the city of Boulder and managed by the Parks and Recreation
Department. It is a city landmark and was placed on the National Register of Historic Places in
1997. Over the past decade, the Parks and Recreation Department and Historic Boulder, Inc.,
have been implementing a Columbia Cemetery Preservation Master Planthanks in large part
to funding from the Colorado Historical Society's State Historical Fund. Each grave marker was
digitally photographed, and more than 100 have received professional restoration work. Several
hundred others have received help from a group of volunteers that comprises the Columbia
Cemetery Conservation Corps.
Members of the Boulder Genealogical Society, in particular Mary McRoberts, have scoured
historic burial ledgers, mortuary documents, obituaries, and court records to compile information
about persons interred in Columbia Cemetery. McRoberts' information helped to make history
come alive for volunteers as she shared the life story of each person whose stone underwent
conservation work. The Boulder Genealogical Society published an eight-volume set entitled,
Columbia Cemetery, Boulder, Colorado, 1870 to the Present. In addition, McRoberts prepared
maps of each of the burial lots in Columbia Cemetery, indicating who had purchased the lots as
well as precisely who is buried within the lot and where. An index of Columbia Cemetery burials
is listed on the Boulder Genealogical Society's Web site (www.rootsweb.com/~bgs).
The Web site project actually began in 2002 when the Boulder Genealogical Society kindly
gave the city permission for use of its Index of Burials and biographical information contained
in the eight volumes. Oracle-based ArcSDE was used to store a cemetery map that was
digitized and registered to the city's aerial photography basemap. Tables were created to hold
each name, biographical sketch, cemetery lot, and grave marker photograph. By linking the
biographical information table to a grave lot feature, Web site visitors are able to query and
display biographical information with ArcIMS software's query server. Custom JavaScript was
used to send XML requests to ArcIMS software's query server and then parse the responses
to generate attractive Web pages presenting maps, scanned records, photographs of grave
markers, lists of all people buried in a particular cemetery lot, and biographical information.
What does the future hold for Columbia Cemetery? Hopefully, grant monies will continue
to provide funding for ongoing preservation of the burial ground, and grave markers will be
carefully repaired or restored one by one. The Columbia Cemetery Conservation Corps has
been working in the graveyard on Saturdays for five summers and shows no sign of stopping.
New ordinances are in place to help protect the graveyard, and the community enjoys strolling
the grounds, picnicking, and studying the fascinating old markers. Tours organized each year
JULY 2009
10
GIS IN ARCHAEOLOGY
by the Parks and Recreation Department and Historic Boulder, Inc., are hugely successful.
Educating the public about Columbia and old cemeteries in general is considered to be the
best tool for fostering the appreciation and respect that will ultimately encourage people to help
protect these cultural treasures. And with Columbia Cemetery information now available to Web
users across the globe, perhaps additional information will come to light as the site is visited by
persons who have knowledge of Columbia Cemetery "residents."
11
WWW.ESRI.COM
An Aztec conquest record from the Codex Mendoza showing place-name glyphs.
The seated figure shown toward the bottom left is the Aztec king Axayacatl who
ruled from ca. 1469 to 1481. The figures surrounding him are the place-name
glyphs for towns he conquered and subordinated to the Aztec empire including
a figure at the top center representing a larger town named Tlatilulco. Tlatilulco's
lord Moquihuix was conquered by Axayacatl and is shown being thrown down
from the main temple of his town, a symbol of total defeat. The scale of the
region is approximately 40 miles (64 km).
13
WWW.ESRI.COM
The Aztec culture developed sophisticated political structures composed of complex hierarchical
systems of social authority, economic infrastructure, and interprovincial tribute. Following
conquest by the Spanish in 1521, elements of these systems formed the base for colonial social
and political structures, many of which persist as modern regional divisions.
Previous research on reconstructing Aztec political geographies typically involved identifying
modern place-names that have indigenous origins, matching these names to historically
referenced locations, and representing these locations with generalized cartographic displays.
This article describes a different approach using GIS. This approach matches uncertain
locations from the past to precisely known present-day locations. The GIS approach was
decisive in developing detailed cartographic summaries of social, political, and temporal
phenomena of interest. Integration with geodatabase technology provided the ability to leverage
tabular records derived from historical research at multiple spatial scales. With a geospatial
database at the core of the investigation, it was possible to process, visualize, and examine
data quickly and efficiently.
The Historical
Record
Investigating geospatial historical records relating to the Aztec rule in central Mexico is
challenging. The documentary evidence available covers a period approximately 20 to 50
years after the Spanish conquered the region. Investigators used Aztec and Spanish evidence,
together with historical and geographic expertise, to look back to the preconquest period.
Spaniards were interested in preserving the Aztecs' political and tribute structures to collect
taxes and develop their own political and ecclesiastical systems within existing native systems.
The Codex Mendoza [a document created to describe Spain's newly conquered lands to
her king, Charles V] is one of the most important primary source documents about ancient
central Mexico and Aztec political geographies. The Codex Mendoza provides place-specific
information describing conquest histories, daily life, and the tributes paid to the central empire.
Images in this document function as a map of historical events. These events are depicted in a
non-European cartographic manner and demonstrate some of the challenges in creating GIS
representations based on the historical record.
Historical
Database Schema
Development and
Implementation
JULY 2009
A computational system was developed to capture, store, and represent historic information and
help understand issues associated with pre-conquest central Mexico. A historical database was
developed to support two core tasks: creating a repository for historic information that would
provide query and analysis tools for nonspatial information retrieval and generating a database
schema that would provide attribute base information for GIS visualization.
14
GIS IN ARCHAEOLOGY
15
WWW.ESRI.COM
Using a GIS-Based
Approach
ArcGIS 9 ArcInfo was used for performing analysis, processing data efficiently, and querying
data structures on the fly. A place-name matching algorithm was developed to match historic
place references to modern place-name points using GIS. The algorithm accepts a time period
as an initial input parameter through a simple GUI. The algorithm uses a series of processing
steps that determine the match confidence. These steps proceed from human-focused data
matches (e.g., a modern town name for an ancient site has been established through careful
historic research) to computational string comparison matches that evaluate multiple partial
matches.
National Geospatial-Intelligence Agency (NGA) GEOnet Names Server (GNS) supplied placename points that were essential to the process of matching ancient place-names to modern
place-names. GNS provides access to databases of geographic feature names maintained by
NGA and the United States Board on Geographic Names. These databases supply 5.5 million
names for approximately four million geographic features worldwide excluding the United States
and Antarctica. Data can be downloaded in CSV file format with x,y values stored in a latitude/
longitude (geographic) coordinate system in World Geodetic System of 1984 (WGS84).
The output table from the name-matching algorithm was joined to the NGA place-name layer,
creating a table that contained data keys for both the modern place-name layer and historical
location records. With these links established, attribute information derived from historic
research could be rendered. Depending on the type of analysis, historical records that could not
be located were dropped from the join.
Once the historic attribute information was joined to the modern place-name point layer, attribute
renderings were created based on specific research needs. The ArcGIS Maplex extension was
used for developing automated label schemes that intelligently placed ancient and modern
place-name labels for each location. Custom label expressions of contrasting styles for modern
and historical names were derived from a single data table. The following code is an example of
a label expression that displays the modern place name in Arial, and the ancient name is shown
in parentheses in a smaller, italicized font.
[ANCIENT_PLACENAME] & VbCrLf & " _ " & "(" &
[MODERN_PLACENAME] & ")" & ""
JULY 2009
16
GIS IN ARCHAEOLOGY
Map-based annotation layers were used extensively for fine tuning of label placement after the
initial label layout was generated. Given the large number of annotations, the ability to make
manual graphic adjustments saved many hours of map production time.
OBJECT_ID
FK_PHPR
ANCIENT
PLACENAME
2464
1365
Comalco
0
0
1148
1346
Cotlaxticpac
Cuauhtenco
3019
1274
Cuahuacan
MODERN
PLACENAME
San Jose
Comalco
MATCH
CONFIDENCE
FOUND_FROM_LIKE
NO_MATCH
NO_MATCH
Santa
Maria
Magdalena
Cahuacan
PLACE CONFIRMED
BY RESEARCH
An important feature of Aztec political geographies is how places interacted through political,
social, and economic systems. Conceptual representations of place relationships were based
on temporal period and visualized using line-direction symbols. Interpretation of the historical
record ranged from clear indications of place interactions confirmed by explicit references to
place-names to generalized, inferred, or uncertain references concerning place interactions. A
directional arrow showed the interaction between Aztec capitals and subjects. The certainty or
uncertainty of an interaction was recorded as an attribute as well as the name of the TO and
FROM points of the interaction.
To effectively show these interactions, an automated spider diagram creation procedure was
developed using ArcObjects. This procedure helped draw lines of interaction by querying
capital/subject relationships stored in the historical database. This saved time and provided
an automated, exploratory method for finding sites when visual inspection was no longer
effective such as when multiple locations with the same name were located near each other.
Approximately 25 percent of the time, computer-generated connections were wrong and
subsequently dropped from the historic database, but sometimes key links were found.
17
WWW.ESRI.COM
Future Research
As the analysis moves further back in time, the historical record no longer provides evidence
and archaeological investigation will be required. Regional-level place interaction analysis can
determine which sites should receive a higher priority when developing research agendas for
archaeological field verification of locations.
Improved modern place-name database development, another area for future research, could
include a systematic comparison of NGA place-name points with annotation labels shown
on 1:50,000-scale quadrangle maps from the Instituto Nacional de Estadstica Geografa e
Informtica (INEGI). In less populated places, place-names don't exist in the NGA database but
are shown on the INEGI maps. Locating a broader range of modern place-names makes more
effective historic reconstructions possible. Finding a Mexican basemap series that predates the
1988 INEGI series would aid historic name identification by providing additional information on
variations in place-name spelling.
Summary
GIS can aid interdisciplinary research in history, archaeology, and geography. The use of GIS
and relational database management technology and datasets helped manage information and
increased understanding of where ancient Aztec places were located and how they interacted
with each other over time. Through improved data structuring, visualization, and analysis, GIS
has also helped manage issues of uncertainty that exist in the historical record.
References
Berdan, Frances, and Patricia Rieff Anawalt, Editors. 1997. The Essential Codex Mendoza. Los
Angeles: University of California Press.
National Geospatial-Intelligence Agency: GEOnet Names Server (GNS), earth-info.nga.mil/gns/
html.
Acknowledgments
Portions of this research were funded by master's thesis research awards from the Association
of American Geographers (AAG) Cartography Specialty Group and the Terry G. Jordan-Bychkov
Student Research Award from the AAG Historical Geography Specialty Group.
The author thanks Dr. Michael E. Smith of Arizona State University (formerly of State University
of New York, Albany) for providing the idea for the project, many of the critical datasets used in
the analysis, and general direction for the project. The author would also like to thank
Dr. Alan M. MacEachren and Dr. Deryck W. Holdsworth of Pennsylvania State University,
Department of Geography, for their assistance.
(Reprinted from the JanuaryMarch 2006 issue of ArcUser magazine)
JULY 2009
18
GIS IN ARCHAEOLOGY
19
WWW.ESRI.COM
The team of investigators that conducted the survey are part of a National Science Foundation
(NSF) grant to examine the nonelite residential occupancy of the site, a research approach
that purposely de-emphasizes the investigation of so-called "monumental" structures, such as
pyramids, temples, and tombs. A 4.2-square-kilometer systematic survey was conducted to
find clues as to where houses and other nonelite structures were located, and the results of the
survey will be used to prioritize which areas will be excavated in the following field season.
JULY 2009
20
GIS IN ARCHAEOLOGY
ArcGIS Desktop (ArcView) was at the core of the geospatial solution developed for the field
survey. In addition, Garmin eTrex Legend GPS devices were used for field data collection.
These devices are easy to use and rugged; have large storage capacity; were accurate for
the spatial needs of the field survey (within 1 to 5 meters at most control points when tested);
and, at a cost of less than $200 per unit, were within the project budget. In addition, the free
Minnesota Department of Natural Resources (DNR) Garmin Extension for ArcView facilitated
eTrex Legend's GIS functionality.
This DNR program provided the key link to integrating GPS-collected data with ArcView. The
program served as a bridge between the Garmin device and the GIS and allowed waypoints to
be uploaded and downloaded from the GPS device, exported raw GPS data to shapefile and
geodatabase feature classes, and provided functionality for synchronizing digital photographs
with GPS waypoints to create hot link features.
All data collected was stored in a personal geodatabase, which was selected because of
its native Microsoft Access file format (.mdb). This allowed spatial data to be integrated
with nonspatial datasets that were developed in a standard Access database. Access data
entry forms were quickly developed on personal geodatabase tables without any concern of
GIS BEST PRACTICES
21
WWW.ESRI.COM
damaging the spatial component of the data as spatial fields were simply left out of data entry
interfaces. In addition, nonspatial data could be quickly and easily visualized in the GIS using
join operations onto spatial data without any interoperability issues arising from trying to join
data in different file formats.
Surface artifacts collected and mapped in the field provide insight into where subsurface
structures and other cultural features may lie.
The ArcView functionality proved essential to the project. An important data source used in the
field surveys was digital orthophotos purchased from the state of Mexico. When the projection
of these photos was found to not correspond with ground-truth data collected by the GPS
devices, georeferencing functions of the ArcGIS Spatial Analyst extension were used to adjust
the imagery, which after rectification and warping then matched correctly with the GPS data.
JULY 2009
22
GIS IN ARCHAEOLOGY
Because of budgetary constraints, the crews used hard-copy maps in the field to mark relevant
information. A critical component of these maps was Universal Transverse Mercator (UTM)
coordinate grids displayed at multiple intervals. Field crews matched the coordinates from these
grids shown on the paper maps with their real-time position coordinates shown in the GPS. This
allowed the crews to stay oriented in the field and not waste time continually trying to figure out
their position from the paper maps alone. A custom ArcObjects script was created to generate
grid lines at user-defined intervals in a grid feature class layer that was used to generate these
maps.
In conclusion, the use of ArcGIS Desktop coupled with a cost-effective GPS/GIS solution,
ease of use, and integration capabilities of geodatabase technology allowed for the rapid
development of a comprehensive geospatial solution to meet the needs of an archaeological
field survey. By using this approach, financial and temporal constraints of the project were
overcome without compromising the scientific integrity of the overall research goals. Although
applied to archaeology, this type of approach could be used by any researcher who needs an
integrated GIS/GPS/data solution but lacks the financial resources for an optimal solution.
More Information
Acknowledgments
Funding for this project comes from NSF grant #0618462, Urbanization and Empire at the
Aztec-Period Site of Calixtlahuaca, Dr. Michael E. Smith, principal investigator. Additional
funding was provided by Arizona State University. ESRI software was provided through the
ESRI campuswide site license program (www.esri.com/campuswidelicense) at Arizona State
University. The project also thanks the Instituto Nacional de Antropologa e Historia (INAH) and
the state of Mexico for their cooperation.
(Reprinted from the Fall 2006 issue of ArcNews magazine)
23
WWW.ESRI.COM
25
WWW.ESRI.COM
For more than 10 years, the Mid-Pacific Regional Office has been in the process of digitally
capturing archaeological site data. A significant portion of archaeological site information for the
region is in a GIS. This information has been collected from a variety of sources for different
projects. Site information is considered to be sensitive data, so regional staff archaeologists
must review and approve access to cultural resource information.
Managing the digital data became increasingly difficult. The spatial data was stored under
different project folders, and site attributes were stored in different tables in a variety of formats.
For better control of the existing digital records and to avoid duplication of records, the data
is being centralized into a geodatabase. The geodatabase offers advantages in identifying
potential duplicate site locations. It has common table definitions and structures. It permits the
JULY 2009
26
GIS IN ARCHAEOLOGY
application and use of a controlled vocabulary and assists in reporting and tracking the status of
site designations by the appropriate authority. The geodatabase also identifies the location of full
site records and survey reports or links directly to any available digital copies.
The geodatabase was constructed as a personal geodatabase with the ArcGIS Desktop
(ArcEditor) application ArcCatalog. As a management database, a site is represented only
once. The geodatabase model assists in preventing duplicate loading of a site multiple times
from separate project databases. It also prevents reporting of the same site with different site
designations by multiple investigators. Sites and site features are represented as points, lines,
and polygons. Sites that are in close proximity are identified and reviewed in ArcEditor, which is
27
WWW.ESRI.COM
also used for verification and validation of proper location and site name. As the geodatabase
develops, ArcEditor provides the opportunity to set topological rules to assist in ensuring that the
sites and site features are appropriately represented for the different feature types.
The geodatabase model has allowed the development of definitions and domains of valid codes
that are consistent with the federal cultural resources management program. Definitions and
the domain of valid codes led to lively discussions among the archaeologists. This included
the application and meaning of terms such as Poor, Fair, and Good. These discussions also
assisted in identifying components of the geodatabase that the archaeologists needed in
performing their tasks.
Domains assist in standardizing information associated with the sites and provide a controlled
vocabulary for the geodatabase. These domains cover the time period associated with the site,
the organization that has surface jurisdiction for the site, site condition, and collection status.
This assists the staff archaeologists in completing surveys and reports for consultation and site
designations.
The ArcMap application within ArcEditor makes it easy for the archaeologists to craft maps
that meet the needs of the various offices of historic preservation. A major issue for the
archaeologists has been the assignment of the proper site identifier or name. The geodatabase
and ArcEditor have simplified their task in selecting and formatting the name to meet their
needs. It also assists in tracking the status of sites that are currently part of consultations
required under sections of the NHPA. A major task for the archaeologists is tracking areas
covered by any cultural resource surveys. This is a component of the geodatabase.
Access to the actual site record can assist in categorizing sites for analysis and in refining the
domain of site types. Viewing the site record is often preferred by the archaeologists when
selecting subsets of sites for analysis. At this time, only a portion of the site records is in digital
form. The staff members have direct access to these records for review and analysis by linking
to the full site record via the geodatabase. Other records only have major characteristics or
features identified as historic or prehistoric attributes in the geodatabase. Both methods are
used in the GIS by the archaeologists for selecting subsets of sites with spatial or attribute
queries.
(Reprinted from the Spring 2006 issue of ArcNews magazine)
JULY 2009
28
GIS IN ARCHAEOLOGY
29
WWW.ESRI.COM
Second, the construction of the database facilitates quick reference when creating reports on
area impacts, such as environmental assessments, or when making quick decisions on potential
resource impacts, such as in the case of fire-line bulldozing in areas of potential cultural
sensitivity.
In short, data layering using ArcGIS Desktop allows complex corollaries to be drawn between
datasets. For example, when land jurisdiction is in dispute over a located cultural resource,
JULY 2009
30
GIS IN ARCHAEOLOGY
transparent land allocation maps overlaid on site data easily solve disputes. These seemingly
simple ways of working with data will do nothing short of revolutionize the way archaeologists
work with cultural data.
Currently, the resource database has seen limited use within BLM due to reluctance to shift
to digital formats. The California Desert District and Surprise field office were the first to get
100 percent of their data uploaded. The Alturas field office has digitized the majority of its site
information but still has some work to do filling out attribute tables for the information. Alturas
may be able to finish up in another season. Hopefully, 2007 will be the year the rest of the BLM
archaeology program goes digital!
Bradley L. Garrett was an archaeological technician for the United States Bureau of Land
Management, Alturas field office, for the summers of 2005 and 2006.
(Reprinted from the Spring 2007 issue of ArcNews magazine)
31
WWW.ESRI.COM
33
WWW.ESRI.COM
Screenshot of Vermont Archaeological Sensitivity Model (VTASM) map document showing sensitivity surface,
custom toolbox, and a couple model examples.
VTASM was developed by a team of researchers from three organizations: ESRI Business
Partner Earth Analytic, Inc.; UMFARC; and the University of Vermont Consulting Archaeology
Program (UVMCAP). Project funding was provided by the Vermont Agency of Transportation
and the Vermont Division for Historic Preservation. Earth Analytic, Inc., served as the GIS
technical lead for the development and implementation of VTASM. A GIS steering committee
composed of archaeologists from a variety of state and federal agencies and institutions
provided oversight and feedback.
VTASM is implemented with ArcInfo, ArcGIS Spatial Analyst, and ArcGIS 3D Analyst, taking
full advantage of ArcGIS ModelBuilder software and the ArcGIS application ArcToolbox. At
JULY 2009
34
GIS IN ARCHAEOLOGY
the core of the system is a functionally and thematically organized directory structure for GIS
data, documents developed with the ArcInfo application ArcMap, toolboxes, exported maps,
and documentation. The VTASM user interface is an ArcInfo document that points to all
required model inputs and a custom toolbox containing about 20 ArcGIS models: flowchart-like
representations of sequences of GIS data management and analysis processes. The VTASM
toolbox is subdivided into two toolsets: one for data preprocessing and one for statewide
analysis. Geoprocessing environment settings are configured at the level of the toolbox,
simplifying the process of changing default settings (workspace and scratch space locations,
output extent, mask, and cell resolution) for the entire statewide model.
The project database includes statewide wetland and hydrological datasets, including the highresolution (1:5,000) Vermont Hydrographic Dataset (VHD), as well as Soil Survey Geographic
(SSURGO) data for most of the state. A notable data limitation is the absence of 10-meter digital
elevation models (DEMs) for the state, although the model does incorporate lidar-based 8-meter
DEMs for a subset of the project area.
Five major preprocessing models prepare specific datasets for use in the statewide model:
hydrological nodes (confluence and terminus points, collectively referred to as hydronodes),
lidar, floodplain soils, streams, and wetlands. For example, one of these models draws on
outputs from four watershed-specific hydronode preprocessing models applied to each of the
17 Vermont watersheds (USGS HUC8). Another preprocessing model converts multiple CAD
point datasets into a triangular irregular network (TIN), then converts the TIN into an eight-meter
resolution raster.
35
WWW.ESRI.COM
Putting the model to use, this shows the VTASM surface overlaid on the NRCS Vermont OnePlan map service.
The statewide analysis toolset consists of 11 environmental component models (ECMs) that are
combined in a composite archaeological sensitivity model. Each ECM yields a statewide
10-meter resolution raster with binary cell values. In each raster, cells meeting model criteria are
assigned a value of one and remaining cells get values of zero.
Six ECM models assign archaeological sensitivity scores to buffer zones associated with
specific water-related features: drainages, water bodies, wetlands, stream confluences,
stream-water body confluences, heads of draws, and waterfalls. For example, the Drainage
Proximity ECM generates a raster buffer zone of 180 meters around the preprocessed
statewide VHD drainages. All cells within 180 meters of streams are assigned a value of one
in the output raster. Given the large size of input datasets, the use of raster-based buffering
JULY 2009
36
GIS IN ARCHAEOLOGY
37
WWW.ESRI.COM
Santa Cruz County is located in southeastern Arizona adjacent to the United States-Mexico
border. It comprises an area of approximately 3,200 square kilometers (km2) of Basin and
Range province topography with elevations ranging from 900 to 2,880 meters (m) above sea
level. The Santa Cruz River, the area's dominant geographic feature, flows through the border
and is flanked by the Santa Rita and Patagonia mountains. Except in the mountains, where
evergreen woodland exists, the Sonoran and Chihuahuan desert environment dominates.
Annual precipitation ranges from 300 millimeters (mm) in the low desert to 900 mm at higher
elevations.
39
WWW.ESRI.COM
Map showing the locations of the archaeological sites and survey projects used in the model.
Culturally the area has been the location of human habitation since the early stages of New
World occupation up to 13,000 years ago. It is also the location of early agricultural sites and a
setting for the introduction of domesticates to the region as much as 4,000 years ago. During
the last centuries before European contact, the area was on the border between the Hohokam
and Trincheras culture areas, and when the first European explorers arrived, the area was
occupied by native Piman groups, whose descendents still live here. Overall, there is good
evidence that some parts of the area were occupied fairly consistently for several thousand
years, and many areas of occupation in earlier times are still important in contemporary land
use.
JULY 2009
40
GIS IN ARCHAEOLOGY
The Project
A goal in creating a sensitivity map of this area was to illustrate the integration of natural and
cultural resources and provide a sense of their value in a set of interpretive themes that highlight
the area's distinct heritage. Of these 10 themes focused on the unique biological and cultural
qualities of the area, twoNative American Lifeways (11,000 B.C. to present) and Desert
Farming (2,000 B.C. to present)are related to prehistoric settlement and land use. Most
themes also emphasize various aspects of life along the river oases that are so prominent in
this desert environment and provide the unifying principle for NHA. Demonstrating the
archaeological aspects of this relationship between land use and the Santa Cruz River system
was a key focus of the analysis.
41
WWW.ESRI.COM
The larger goal in obtaining NHA designation was the development of heritage and nature
tourism in the area. Over the first 10 years, the impact of increased tourism resulting from the
NHA designation is estimated at approximately $1.8 billion in revenue and 40,000 new jobs. To
most effectively manage this development and its impacts, it is necessary to illustrate where
cultural resources are concentrated, their relationship to other resources, and how they will be
affected by increased tourism activity.
The statistical technique used for this predictive model was the logistic regression technique,
described by K. L. Kvamme in a 1983 article (see References), which provides output ranging
from 01 that can be interpreted as a probability. Logistic regression is suitable for use with
binary-dependent variables and a range of independent types including categorical variables
common in environmental studies. For site locations, a set of 160 pre-European contact
habitation sites was obtained from AZSITE. These sites were chosen because they were likely
to represent the broadest range of activities relevant to the interpretive themes and were most
representative of past land-use and cultural resources that need protection.
For nonsite locations, a set of archaeological survey polygons was obtained from AZSITE
comprising 148 projects covering more than 70 km2 and representing approximately 2 percent
of the area of Santa Cruz County. A nonsite was defined as any area where survey work did
not reveal the presence of archaeological resources. Nonsite areas were converted to point
locations to create a dataset with attribute association comparable to site locations. This
process resulted in the identification of 310 nonsite locations distributed broadly across Santa
Cruz County.
Obtaining and developing useful environmental data can be the most time-consuming and
costly aspect of a predictive modeling project. Project constraints required the use of available
data that was typically collected for quite different purposes and recorded at a scale that might
not be appropriate for modeling some aspects of prehistoric land use. Consequently, it was
necessary to derive meaningful attributes from this data while still discriminating statistically
useful variation.
Looking at Water
Availability
JULY 2009
One especially important consideration in site location is difficult to address with standard
hydrographic data. This is the availability of water. In the deserts of the Southwest, this was an
important consideration for prehistoric settlers. It is difficult to identify this concern with current
data in a way that reflects actual availability. Hydrographic data available from government
sources typically does not adequately indicate subtle variations in water availability in the desert
nor does this data address differences between current and past conditions. Simple distinctions
42
GIS IN ARCHAEOLOGY
between perennial and ephemeral streams, or methods of identifying stream order, do little to
indicate the actual quantity and timing of water availability that are critical to human uses.
Furthermore, surface water availability has changed greatly over time, particularly in the last
century, because modern uses have affected flow characteristics and the water table.
Map of the results of the BIC approximation model showing areas of archaeological sensitivity in Santa Cruz County.
For example, in the present analysis, only two small segments of the many streams in the area
were identified as perennial; the rest were considered ephemeral. However, a more extensive
perennial flow in the larger streams of this region has been historically well documented. In
addition, the mean distance to any stream from sites used in these analyses was 294 m,
compared to 303 m for nonsite locations. Such a small difference in distance to ephemeral
water sources hardly reflects its importance in this desert region and would provide little
GIS BEST PRACTICES
43
WWW.ESRI.COM
JULY 2009
Attributes of environmental variables were assigned to the collection of sites and nonsites by
location to produce a table for statistical analysis. An important problem in multivariate statistical
models is how to determine which variables will provide the best model. Some discussion in the
archaeological literature suggests using stepwise procedures, adding or removing variables
based on a predetermined significance value. This procedure has been criticized, however, and
the Bayesian Information Criteria (BIC) approximation (Raftery 1995) offered a useful alternative
for distinguishing the best set of variables for creating an efficient predictive model. For this
project, BIC approximation indicated that a set of variables including flow accumulation,
elevation, distance to springs, soils, and vegetation provided the best model. Further tests of
this model against models suggested by stepwise procedures indicate superior performance of
the BIC model.
44
GIS IN ARCHAEOLOGY
Results and
Discussion
The resulting predictive model provided mean probability estimates of .97 for site locations and
.03 for nonsite locations, indicating strong discrimination between location types. Reclassifying
the probability map, so values above .5 are positive predictions and values below are negative,
correctly predicted 98 percent of the sites and 98 percent of the nonsite locations. A more
efficient model correctly predicted at least 90 percent of the site locations by reclassifying
positive and negative predictions at a .96 cutpoint that focused attention on only 21.4 percent of
the total area as archaeologically sensitive. This model represents a 76 percent gain in
efficiency in identifying areas of interest for cultural resources. Moreover, it is in stronger
agreement with other models of natural resources, such as biodiversity, that are also an
important part of the NHA.
Despite the present need for a single model, it is essential to consider modeling as an ongoing,
iterative process. The current project is only a feasibility study and will be followed by continued
efforts as the NHA designation advances. Ultimately, more detailed management plans may
require more elaborate modeling efforts and the consideration of new variables as necessary to
clarify understanding of particular problems and relationships.
Also, understanding the archaeology of this region will certainly improve as more research is
conducted in coming years. As development expands greatly in the area, it will change the
articulation of past and present interests. The present model cannot be considered a final
word on archaeological sensitivity in Santa Cruz County. Rather, these efforts and the lessons
learned will, it is hoped, serve as a productive foundation for continued work. It is encouraging
that this initial project has offered a useful model and numerous valuable insights into the
modeling process. For a more detailed description of this project, visit www.cdarc.org/pages/
national_heritage/areas/scnha.php.
Acknowledgments
References
The authors thank William H. Doelle for the support necessary to conduct this research and
pursue the establishment of the Santa Cruz Valley National Heritage Area. They also thank Chip
Colwell-Chanthaphonh, Jeffery J. Clark, Patrick D. Lyons, Miller McPherson, Anna A. Neuzil,
and Linda J. Pierce for their thoughtful comments on this research.
Kvamme, K. L., 1983. "Computer Processing Techniques for Regional Modeling of
Archaeological Site Locations," Advances in Computer Archaeology, 1:2652.
Raftery, A. E., 1995. "Bayesian Model Selection in Social Research," Sociological Methodology,
25:111163.
45
WWW.ESRI.COM
J. Brett Hill holds a doctorate in anthropology from Arizona State University. He is a preservation
archaeologist at the Center for Desert Archaeology in Tucson, Arizona, and a visiting assistant
professor of anthropology at Hendrix College in Conway, Arizona. His interests are in GIS,
human ecology, and land use in desert environments.
Mathew Devitt is a master's student in anthropology at the University of New Mexico,
Albuquerque. His interests are in ceramic analyses and the late prehistory of the southwestern
United States and northwestern Mexico.
Marina Sergeyeva is a master's student in anthropology at the University of Illinois,
Urbana-Champaign. Her interests are in GIS and Paleolithic archaeology in Russia.
(Reprinted from the AprilJune 2007 issue of ArcUser magazine)
JULY 2009
46
GIS IN ARCHAEOLOGY
ESRI 380 New York Street Redlands, CA 92373-8100 909-793-2853 FAX 909-793-5953 www.esri.com